An Independent Component Analysis(ICA) Circuit Design with Its Applications to Signal Separation

• Main Authors: Shin-Hung Tsai, 蔡信宏
• Other Authors: Szi-Wen Chen
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/56303234795705518043

碩士 === 長庚大學 === 電子工程研究所 === 95 === Independent Component Analysis (ICA) is known as a novel algorithm which is developed to solve the blind signal source (BBS) problems. The characteristic of the algorithm is that it allows a direct separation for a number of mixed signals with unknown coefficients of mixture. Therefore, the ICA algorithm becomes important in the field of the digital signal processing (DSP). This thesis research aims to design and implement a real-time hardware architecture of the extend infomax ICA algorithm that can separate the super-Gaussian signal sources using the Verolig Hardware Description Language (Verilog HDL). To validate the architecture, a number of mixed measured ECG signals were selected in this study for software and hardware simulations. According to our numerical and physical experimental results, the real-time architecture may be successfully applied to separating mixed medical signals into independent sources. Moreover, the designed architecture may be further integrated into an embedded system.